基于条件归一化网络的艺术字体多风格融合技术研究  

Research on Artistic Font Multi-style Fusion Technology Based on Conditional Normalization Network

作　　者：陈舒祥 王江江 战国栋 CHEN Shuxiang;WANG Jiangjiang;ZHAN Guodong(School of Information and Communication Engineering,Dalian Minzu University,Dalian Liaoning 116605,China;School of Computer Science and Engineering;School of Design,Dalian Minzu University,Dalian Liaoning 116605,China;Dalian Chinese Font Design Technology Innovation Center,Dalian Minzu University,Dalian Liaoning 116605,China)

机构地区：[1]大连民族大学信息与通信工程学院,辽宁大连116605 [2]大连民族大学计算机科学与工程学院,辽宁大连116605 [3]大连民族大学设计学院,辽宁大连116605 [4]大连民族大学大连市汉字计算机字库设计技术创新中心,辽宁大连116605

出　　处：《大连民族大学学报》2024年第3期248-253,共6页Journal of Dalian Minzu University

基　　金：辽宁省自然科学基金项目(2020-MZLH-19);贵州省科技支撑计划项目(2021-534)。

摘　　要：针对艺术字体风格迁移中风格单一性以及多风格字体融合时出现的风格单一、风格溢出和残余黑点的问题,提出了一种基于条件归一化网络的多风格字体图像融合方法。该方法采用了一个字体多风格融合模型,能根据不同的风格参数调整预训练的VGG19Net网络的输出。通过将内容图与多个风格字体图像分别提取特征,并使用一个新的风格损失函数来度量内容图与风格字体图像之间的差异,将多个风格特征混合,生成新的艺术风格字体图像。该方法的关键在于卷积层和条件归一化层,能够实现多风格的迁移和融合。实验结果表明:与其他两种方法相比,内容相似度平均提高了1.2%,风格相似度平均提高了2.7%。In order to solve the problems of style uniformity,style overflow and residual black dots in multi-style font fusion,a method of multi-style font image fusion based on conditional normalization network is proposed.This method employs a font multi-style fusion model that adjusts the output of the pre-trained VGG19Net based on different style parameters.By extracting features from content images and multiple style font images separately,and using a new style loss function to measure the differences between content images and style font images,multiple style features are blended to generate new artistic font style images.The key components of this method are the convolutional layers and conditional normalization layers,which enable the transfer and fusion of multiple styles.Experimental results show that,compared to other two methods,the proposed method achieves an average improvement of 1.2%in similarity of content and 2.7%in similarity of style.

关 键 词：风格迁移 多风格融合 条件归一化 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]